GB2608675A

GB2608675A - System and method for the measurement of velocity

Info

Publication number: GB2608675A
Application number: GB2206354.9A
Authority: GB
Inventors: Schlegel Ralf
Original assignee: Individual
Current assignee: Individual
Priority date: 2021-04-30
Filing date: 2022-04-29
Publication date: 2023-01-11
Anticipated expiration: 2042-04-29
Also published as: GB2608675B; DE102021111221A1; GB202206354D0; DE102021111221B4

Abstract

The invention relates to a system and a method for determining the velocity of objects passing through the system through two fields of view of the system. A system in accordance with the invention for determining the velocity of passing objects (O) comprises a means for control (40) of a first optical sensor (10) and a second optical sensor (20), wherein the means for control (40) continuously alternates between an adjustment phase and a measurement phase, wherein in the measurement phase, with gains V1set, V2set, previously adjusted in the adjustment phase, of first and second transimpedance amplifiers (14, 24), connected respectively to the first and second optical sensors (10, 20), for determining the velocity of passing objects (O), a measurement of the time profiles of the brightness in the first and second fields of view (12, 22) is performed .

Description

System and method for the measurement of velocity The invention relates to a system and a method for the measurement of velocity, in particular a system and a method for determining the velocity of objects passing through the system, through two fields of view of the system. The invention enables the detection and evaluation of time-displaced light signal patterns with the aim of determining a velocity of object passages, previously identified or classified in the light patterns, from the time displacement of at least two light signal patterns, and a known field of view separation distance, under variable ambient conditions.

A practical example of the application of corresponding systems and methods is the detection and provision of data for traffic planning and analysis. There therefore already exists a diverse selection of methods and procedures for the measurement of velocities, especially of passing motor vehicles on carriageways and other traffic routes.

These include those using induction measurement loops (e.g. CH 0 201 126 A, DE 01 574 053 B); radar (e.g. DE 03 727 562 C2, DE 3 902 582 A1, EP 033 509 A2), or laser-based methods (Doppler effect); those using pressure hoses that are laid across the carriageway so that they can be driven over; or other approaches using acoustic (e.g. DE 10 2015 120 533 A1, magnetic (Hall effect, e.g. DE 3 830 598 A1) and optical sensors (e.g. DE 196 004 338 B4, DE 3 902 582 A1, EP 033 509 A2). The attenuation of external radio fields by passing objects (e.g. B. DE 10 2018 109 990 A1) is also called upon for the determination of the velocity of motor vehicles. There are also methods based on mobile radio equipment with transmitter units integrated in the vehicles (e.g. DE 19 604 084 Al).

However, optical methods are particularly preferred by virtue of their simplicity. These include, for example, video camera-based systems (e.g. DE 29 809 288 U1, EP 0 651 364 Al), systems using light gates with at least one transmitter/receiver combination on different sides of the carriageway, and also single-side sensor systems. The latter can be subdivided into active and passive approaches.

Active methods themselves emit light of a specific wavelength, which is sent back to a receiver, usually installed in the same housing, by reflection from an opposite side of the carriageway, or from a target object passing at right angles to the receiver. The light in question can be adjusted in its properties in various ways, e.g. the use of specific wavelengths, as well as pulsed or frequency-modulated light, is of known prior art. Passive methods, on the other hand, use just ambient light. However, passive methods are generally considered to be sensitive to interference, which is why active systems are usually preferred (e.g. EP 2 503 342 B1).

The basic operating principle of optical side sensors is to determine the time displacement At of signal profiles when an object passes between at least two marking lines or sensors at a known separation distance As, and to calculate the velocity v = As/At from this data. In the case of reflective light gates, the interruption of the light signal usually serves as a trigger event for a start/stop measurement of the time interval for the passage of an object. This is more difficult in the case of single-side sensors, i.e. in the case of systems that are only arranged on one side of the object passage. Here, light signal patterns at right angles to the carriageway are recorded almost synchronously by at least two sensors as soon as an object passes through the monitored field of the system. Measurements are therefore usually initiated when triggered by a trigger signal as a result of the passage of an object (e.g. EP 0 877 253 Al).

With recordings of this type, the data must then be compared with each other in order to determine the time difference. Correlation methods are typically used to compare such data sets, and to determine the time difference (e.g. DE 3 830 598 A1, EP 0 357 893 A2, DE 3 902 582 A1, DE 10 2015 120 533 Al). However, the quality of the result from correlation methods depends essentially on the composition of the signal profiles. The best results are achieved when the underlying data are pseudo-stochastic in nature (cf. EP 0 877 253 Al).

Experience shows that the brightness fluctuations at right angles to the carriageway, measured with single-side sensors, are relatively large when a target object passes. The signal profiles are therefore rather untypical for the field of application of correlation methods. Consequently, greater deviations from the true velocities are to be anticipated in the measured result. In particular, gradients in the data set worsen the accuracy of such correlation measurements. Methods of this type can therefore only be performed on short data segments of the passing measurement object, which must then be available with a high point density, which in turn requires increased measurement complexity. Another disadvantage is that short pseudo-stochastic data set segments cannot really be recognised by external observers as being plausibly assigned to the passage of a target object.

The invention is now based on the object of overcoming, or at least reducing, the problems of the prior art, and providing a simple and compact, single-side operating, passive system, and an associated method, for determining the velocity of an object passing through two fields of view of the system, which offer a higher accuracy and reliability than comparable systems and methods.

In particular, the invention should also enable an autonomous long-term monitoring of, for example, the daily profile of traffic flows without further manual interventions or adjustments, and should be able to analyse, evaluate and present the measurement results directly on site. Here the measurement data should be processed in the system in a way that is comprehensible and verifiable for external observers, and in particular it should be possible for external observers to recognise this data as being plausibly associated with the passing of a target object. In particular, a system and a method for the detection of, and criteria for the evaluation of, light signal patterns for optically-passive velocity measurement systems based on single-side sensors are to be specified. The objective is thus to provide a system and a method, which can determine time difference values for single-side sensor systems with an acceptable level of technical effort, without the disadvantages cited in the prior art for corresponding correlation methods.

The object in accordance with the invention is achieved according to the independent Claims 1 and 7. Preferred developments are the subject matter of the referenced subsidiary claims.

A first aspect of the invention relates to a system for determining the velocity of objects passing through two fields of view of the system, comprising a first optical sensor, set up for a directional detection of brightness values in a first field of view, wherein a signal output of the first optical sensor is connected to a first transimpedance amplifier with variable gain VI"," set up so as to produce an output signal I,, proportional to a brightness in the first field of view; a second optical sensor set up for a directional detection of brightness values in a second field of view, wherein the first field of view and the second field of view are aligned essentially parallel to each other, and at a fixed separation distance from each other, wherein a signal output of the second optical sensor is connected to a second transimpedance amplifier with variable gain V2var, set up so as to generate an output signal 12, proportional to a brightness in the second field of view; a means for control of the first and the second optical sensors, wherein the means for control continuously alternates between an adjustment phase and a measurement phase, wherein, in the adjustment phase, an adjustment takes place of the gains Vlvar, V2var, of the first and the second transimpedance amplifier respectively, into a defined preferred operating range, corresponding to a current brightness in the first and second fields of view, wherein in the subsequent measurement phase, with the gains V1set, V2se1 of the first and second transimpedance amplifiers, adjusted in advance, a measurement is performed of the time profiles of the brightness in the first and second fields of view, so as to determine the velocity of passing objects.

Optical sensors can preferably take the form of photodiodes, but can also take the form of single or multi-channel detectors, in particular as line or array detectors. In the case of multi-channel detectors, a single channel can be used for the measurements. Two or more channels can also -4 -be bundled for a corresponding single-channel detection. The field-of-view (FoV) of an optical sensor describes the spatial region from which light can fall onto the optical sensor for measuring the brightness in this region. Thus, all light rays incident from this spatial region are added up to form a measured total brightness within the field of view. Typically, the field of view expands with increasing distance from the active surface of the optical sensor, so that, despite small active surfaces of the optical sensors, the brightness can still be integrated over relatively large spatial volumes. However, since the brightness of a light source decreases with the square of increasing distance from the latter, the depth of view (DoV) is usually limited to a few metres forward.

The field of view detected by the optical sensor is preferably spatially narrowed, for example by a thick aperture in front of the optical sensor, or with the aid of a lens, or by means of an optical imaging system. A lens, or an optical imaging system, can also preferably be used to adjust a preferred depth range within the field of view of the optical sensor, for example so as to optimise the brightness detection onto a specific distance between the optical sensor and a specific carriageway. Depending on the aperture angle of the respective fields of view, and the separation distance between the optical sensors, the respective effective fields of view can also partially overlap in depth.

The axis of the field of view is generally at right angles to the active surface of the optical sensor. However, its orientation may also be tilted with respect to the active surface of the optical sensor by means of a lens or an imaging system arranged in front of the optical sensor. Regardless of the exact shape and orientation of the individual fields of view, a largely parallel alignment of the first field of view and the second field of view can be assumed, in particular if, by simply displacing one of the two fields of view, one of the two fields of view is completely contained within the other field of view. If the shapes and orientations of the individual fields of view are identical, they will then overlap each other accordingly. The separation distance between the fields of view can be assumed to be the separation distance between the respective beam axes of the fields of view. With a parallel alignment of the optical sensors, and without any tilt or offset due to additional optical components, the separation distance thus generally corresponds to the centre-to-centre distance of the active surfaces of the optical sensors, and approximately to a general definition of the separation distance between the two optical sensors.

A signal output of each optical sensor is connected to a transimpedance amplifier with an adjustable gain VIvar, V2var. The transimpedance amplifiers generate an output signal Ii, 12, which is proportional to the brightness in the respective field of view. By virtue of the adjustable gain Vlvar, V2var, the optical sensors can be optimally adjusted to different light conditions in the environment, in the case of solar radiation in particular, to an irradiance that alters frequently -5 -during the course of the day. Here a transimpedance amplifier with an adjustable gain Vivar, V2var is generally understood to be any current-controlled voltage source that is set up as a current-voltage converter (1-V converter) so as to generate an output signal 11, 12 proportional to the brightness in the respective field of view.

The operating range of an amplifier is generally understood to be the range within the associated characteristic field, within which a near linearity between an input signal and the associated amplified output signal is ensured for all permissible or assumed operating states. The optimum operating range is understood to be in particular that sub-range in which a sufficiently high gain can be anticipated, at the same time with low noise, and a low probability of any overload of the amplifier output. The location of the optimal operating range in the family of characteristics depends on a multitude of factors, and can be determined and defined individually for each combination of optical sensor and transimpedance amplifier. The preferred operating range can be individually defined for a particular system combination.

The means for control of the first and second optical sensors enables a dynamic adjustment to a varying, that is to say, a temporally altering, environmental brightness. In accordance with the invention, the means for control continuously switches back and forth between an adjustment phase and a measurement phase. In each adjustment phase, the gains Vlvar, V2var of the first and second transimpedance amplifiers are in each case preferably adjusted to a defined preferred operating range corresponding to a current brightness in the first and second fields of view.

The adjustment phases are used exclusively to adjust the system to the current external conditions, that is to say, no velocity measurement is performed within these phases. In a measurement phase following an adjustment phase, the gains V1"t, V2Set of the first and second transimpedance amplifiers, which have previously been adjusted to the current ambient conditions, are used to determine the velocity of passing objects by measuring the time profiles (signal profiles) of the brightness in the first and second fields of view.

A concept of the present invention is thus to enable an adjustment, in particular to a temporally altering ambient brightness, by periodically interrupting the measurement with adjustment phases, in which the gains of the transimpedance amplifiers are adjusted. The actual measurement is performed exclusively in the intervening measurement phases.

The duration of an adjustment phase is preferably shorter than the duration of the subsequent measurement phase. In particular, it is preferable for the duration of the adjustment phase to be less than 10% of the duration of the measurement phase, preferably less than 1% of the duration of the measurement phase, and even more preferably, less than 0.1% of the duration of the measurement phase. The duration of the adjustment phases and the duration of the measurement phases can in each case be constant, wherein, however, an ability to adjust the ratio of the durations can be provided.

Alternatively, the duration of the adjustment phase and the duration of the measurement phase can also be adjusted parametrically during operation, that is to say, during a measurement series consisting of a plurality of successive measurement phases. For example, the ratio of the durations can be adjusted depending on the time of day; a known, assumed, or previously measured, traffic volume; a rate of alteration of the brightness in the first and second fields of view; or can be coupled to external triggers (e.g. traffic light phases, closure times). Adjustability has the advantage that the consequences of the effective dead time of the measurement system are minimised, and the system can be optimally adjusted to different requirements. Otherwise, the interval between the adjustment phases should preferably be selected in such a way that alterations in the brightness in the first and second fields of view can reliably be detected with sufficient accuracy, without unnecessarily increasing the dead time of the measurement series.

The system preferably also comprises a means for evaluation, designed so as to identify or classify, on the basis of analytical and/or statistical methods using defined parameters, an object passage in the time profiles, from the time profiles of the brightness in the first and second fields of view measured in a measurement phase. Since the brightness in the first and second fields of view is continuously recorded during the individual measurement phases, without individual object passages already having been detected during the measurement, or marked and identified in the profiles, the latter must be determined by other means. For this purpose, various analytical and/or statistical methods can be called upon for the evaluation. Here, correlation methods of known prior art can also preferably be applied to the two profiles. In accordance with the invention, however, these are used exclusively for identifying or classifying object passages in the profiles, and not for determining velocities.

The means for evaluation thus has the task of searching for relationships in the time profiles of the measurement signals that are most likely to relate to object passages. In particular, if certain characteristic correlation parameters are known, and with the aid of further parameters (for example, the distance of the system from the preferred axis of the object passages, that is to say, from a carriageway, for example), it is also possible to draw conclusions as to the types of the individual objects identified. This process is called classification, wherein a classification necessarily includes, or presupposes, an identification. An identification, on the other hand, merely means that features indicating object passages are searched for in or between the -7 -individual signal profiles, and that these features are assigned to actual object passages according to certain criteria or, for example, rejected as inadmissible or incorrect for reasons of implausibility (for example, an unrealistic velocity range, an object that is too short or too long).

The analytical and/or statistical methods for identifying or classifying object passages preferably use at least one criterion from the following group: Alteration by a defined amount of the difference signal between the measured time profiles of the brightness in the first and second fields of view; Mean values and/or standard deviations, or quantities calculated therefrom, in the difference signal between the measured time profiles of the current brightness in the first and second fields of view; Differences between arithmetic mean values, the confidence intervals, or the difference of the standard deviations of mean values in the measured time profiles of brightness in the first and second fields of view; A half-width in a correlation function between the measured time profiles of the brightness in the first and second fields of view; A number of maxima and/or minima in a correlation function between the measured profiles of brightness in the first and second fields of view.

In EP 0 042 546 A1, object detection is achieved in that time differences are made with a three-sensor or a three-gate system by comparing the At values between gates 1 and 2 and gates 2 and 3. If the difference between the two times is too great, the measurement is considered to be invalid. In a system in accordance with the invention, this function can already be performed with two optical sensors. by the assessment of features of a correlation function of the two sensor signals. In particular, this can take the form of a cross-correlation between the two sensor signals. However, an autocorrelation of one sensor signal at a time is also quite suitable to enable a valid/invalid assessment on the basis of the half-widths.

If, for example, a spontaneous change in brightness caused by natural effects (for example, by clouds in sunlight) occurs simultaneously on both optical sensors, this results in a correlation function with a larger half-width in comparison to the normal passage of a vehicle. If, on the other hand, an event is only detected on one optical sensor, or if the signals on the sensors are not only displaced in time, but also differ noticeably in their intensity, this can be shown by observing the cross-correlation function (for example, the number of maxima and minima occurring) as well as by statistical methods applied to the light signal profiles. In the latter case, consideration can be given to the variables: mean value, standard deviation, and variables calculated from these, such as the variation coefficients, the deviations of the arithmetic mean, the difference between arithmetic means, confidence intervals or also a normalised standard deviation difference (sddiff) of the standard deviations of the individual signal profiles (first optical sensor: sch, second optical sensor: sd2) in accordance with equation (1): I Sal -Sa2 I (1) Sdayf 0,5.(Sit 1 +SC12).

For an identified or classified object passage, the means for evaluation preferably executes the velocity of the object on the basis of a respective polynomial fitting at the extreme points belonging to the identified or classified object passage in the measured time profiles of the brightness in the first and second fields of view, analytically determines the exact time location of the extreme points by forming the derivative of the respective polynomial fitting, and determines the velocity of the object with the time difference resulting from the location of the extreme points by difference formation, together with the known separation distance between the first and second fields of view. In contrast to the methods based on correlation function calculations in the prior art, the velocity of an object is thus determined directly in both signal profiles by the time distance of the features, in each case belonging to a single object passage, and identified or classified as belonging to each other.

Precisely because the light signal profiles can show clear differences in their respective time profiles, that is to say, the sensor signal is not pseudo-stochastic, the determination of a time difference, that is to say, a time displacement, by way of a cross-correlation is unsatisfactory. It can be illustratively shown that a uniform gradient in both signal profiles leads to erroneous results when using a cross-correlation. For this reason, in the inventive system as presented, maxima or minima deviating from the zero lines of background brightness-corrected differential amplifier signal lines (an adjustment, for example, set appropriately during a preceding adjustment phase by means of a "sample-and-hold" element) are preferably calculated analytically precisely by means of polynomial fittings and their mathematical derivatives.

In particular, the velocity calculation of an inventive velocity measurement system can thus be performed precisely on the basis of polynomial fittings (for example by way of a 3rd degree polynomial) at extreme points in the light signal profile (maxima or minima) of both signals, wherein the exact location is determined analytically in each case by forming the derivative, and the time difference determined therefrom is used, together with the known sensor, or field of view, separation distance, to determine the velocity.

The starting point for this can be the determination of the general maximum or minimum, or a plurality of maxima or minima of the measured value channel, wherein the data set can be run -9 -through at least once, and then searched. The limits of a polynomial fitting can then be set to a certain number of data points before or after this maximum or minimum. After execution of the polynomial fitting, the root mean square (rms) value obtained can be compared with a minimum target value. If the rms value obtained from the polynomial fitting is larger than the target value, the fitting limits to the right and left of the roughly determined maximum or minimum can be reduced, and the polynomial fitting can be executed once again for a better adjustment to the extreme value. The procedure can then be repeated until either the target value is achieved, or a minimum number of data points required for the polynomial fitting is achieved or undershot. This procedure is thus excellently suited for achieving a good match to the polynomial, for extreme values of different shapes. With reference to the acquisition of the measured values, the measuring frequency and signal processing should be selected such that the passage of an object is represented by as smooth and even a profile as possible, in the region of the extreme value (comparable to a low-pass filtered signal), which also represents the change in brightness of the environment during the approach of an object and during its departure from the system.

The mathematical relationships for the precise determination of the extreme values can be illustrated in an exemplary manner by the following three equations: y(x) = a() + aix + a2x2 + a3x3 j(_cti)2 = X1,2 2a, 2 a, a2 y'(x) = 2a2 6a3x1,2 Here, Equation (2) shows a general 3rd degree polynomial as a fitting equation, Equation (3) shows the solutions of the 1st derivative to determine the extreme values, and Equation (4) shows the 2nd derivative to determine the type of extreme value present, and to compare or check with the data set.

The determination of the extreme points in the measured time profiles of the brightness in the first and second fields of view can be performed very quickly, efficiently and precisely by fitting to a polynomial. By this means, the deviations from the real velocity that occur in comparable systems, based on correlation function calculations for determining the velocity, as well as the occurrence of false measurements or detections, can be significantly minimised. In addition, the actual determination of the velocity of the identified or classified objects can also be understood very easily and transparently by external observers. Possible problems for external observers with the comprehensibility of a velocity measurement then arise at most at the level of the identification and classification of individual objects. By using an analytical polynomial-based calculation (2) (3) (4) approach, the method is also suitable for gradient-beset time profiles of brightness in the first and second fields of view, which otherwise lead to more inaccurate and uncertain results when using the usual correlation methods for determining the object velocity.

The advantages of the inventive system lie in particular in a simple evaluation circuit based on standard principles, a constantly high sensitivity, even under differing and time-varying measurement conditions, a high level of plausibility of the measurement data for external observers, as well as an increased precision in the time differences determined, compared to correlation methods of the known prior art, at the same time with an increased robustness with respect to any disturbing factors. Due to its simple and data-saving mode of operation, the measuring system as developed is also suitable for velocity measurements that are compliant with data protection requirements, without any recording of the personal data of road users, from both public and private properties.

The system preferably further comprises a means for location determination, set up so as to determine a tilt of the line connecting the first and second fields of view with respect to a predetermined axis of the object passages, and to transfer this to the means for evaluation as a correction to the separation distance between the first and second fields of view. In particular, microelectromechanical systems (MEMS) can be present on the system for this purpose, and evaluated accordingly. For example, inclination sensors, location sensors, and/or barometric pressure sensors, can be used for this purpose. A determination of the location can also be made with the aid of a satellite-based navigation system (GPS or similar). For example, the means for determining location can also determine the location/tilt of the individual optical sensors with respect to the carriageway, and/or each other. By this means, any incorrect positioning of the inventive velocity measurement system can automatically be detected, and can be compensated for mathematically by means of correction parameters, appropriately stored or determined in the system, when determining the velocity of an object.

The inventive system is preferably designed so that it is suitable for integration into building components or enclosures adjacent to carriageways, as well as for integration within a building behind window glass, or in windows, or in window decorative elements, and thus does not require an additional measuring beam or support/holder. In particular, integration into building components (window frames or measuring bars in the interior, as window decorative elements, on/in pillars or walls) or also, for example, on the front or side faces of vehicles, can thus be undertaken. In autonomous vehicles in particular, a system in accordance with the invention can also be used as a component for environment detection, possibly in conjunction with the data of a LiDAR (light detection and ranging) system, or a camera system.

The system preferably also comprises a means for automatic processing, in particular, of the velocities determined, the times of an object passage, and/or of classification results. In addition to these data, the means for automatic processing can also derive further statistically relevant data such as traffic flow densities, peak traffic times, congestion forecasts or similar. The result of automatic processing can be made available to a system operator in electronic or other form as an evaluation report, immediately after a series of measurements. For remote monitoring, a system in accordance with the invention can be equipped with an appropriate communications device (for example, by way of LAN, WiFi, Bluetooth, NFC). The evaluation report can be sent to a system operator, in particular in a cloud-based manner, or by email.

Another aspect of the invention relates to an appropriate method for determining the velocity of objects passing through two fields of view of the system, comprising the provision of a system in accordance with the invention for determining the velocity of objects; and a continuous alternation between an adjustment phase and a measurement phase of the system, wherein in the adjustment phase the gains Vivar, V2"ar of the first and second transimpedance amplifiers are in each case adjusted to a defined preferred operating range, corresponding to a current brightness in the first and second fields of view, wherein in the subsequent measurement phase, with the previously adjusted gains Visei, Vzset of the first and second transimpedance amplifiers, a continuous measurement of the profiles of the current brightness in the first and second fields of view is performed, for determining the velocity on the basis of identified or classified object passages.

The method is directed to the application of a system in accordance with the invention, and therefore relates directly to the functions of the individual features. The parts of the description relating to these features of the system therefore also apply directly to the method. This also applies to the individual forms of embodiment described, including the associated examples, and the advantages cited in each case.

The invention thus relates to a system and a method for the detection of, and criteria for the evaluation of, light signal patterns for optically-passive velocity measurement systems based on single-side sensors. In one form of embodiment, an optical sensor (for example, a photodiode), in combination with a transimpedance amplifier with an adjustable sensitivity, or gain is used for measurement. The mode of operation of the detection of light signal patterns is performed discontinuously, within an alignment phase and a measurement phase, wherein the duration of the alignment phases is preferably small compared to that of the measurement phase. The light signal patterns in the time profiles can subsequently be obtained by forming the difference between a reference signal determined in the calibration phase by way of a "sample-and-hold" element, and a light signal current at the time of measurement.

A means for evaluation can then evaluate the signal profiles of at least two optical sensors, preferably on the basis of the half-width of cross-correlation or autocorrelation functions, the normalised standard deviation, the coefficient of variation, and/or by counting maxima and minima of the cross-correlation function to assess whether a real target object has passed the sensors. The means for evaluation can furthermore analytically determine the time difference, at a time-displaced extreme value in both light signal profiles, by way of a polynomial and its derivatives. By using an analytical polynomial-based calculation approach, the method is also suitable for gradient-beset light signal patterns, which lead to more inaccurate and uncertain results in the determination of velocity with conventional correlation methods.

Further preferred configurations of the invention ensue from the other features cited in the subsidiary claims.

The various forms of embodiment of the invention cited in this application can advantageously be combined with each other, unless otherwise stated in an individual case.

In what follows the invention is explained in terms of examples of embodiment with reference to the accompanying drawings. Here: Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 shows a schematic illustration of a system in accordance with the invention for determining the velocity of passing objects; shows a schematic illustration of an optical sensor with a transimpedance amplifier (TIA), a "sample-and-hold" (S&H) element, and an instrumentation amplifier (INA); shows an illustration of the relative frequency of half-widths of the cross-correlation function using an exemplary data set; shows normalised cross-correlation functions of various signal profiles (a-c) and corresponding signal profiles (d-f); and shows a schematic illustration of a polynomial fitting of extreme values.

Figure 1 shows a schematic illustration of a system in accordance with the invention for determining the velocity of passing objects 0. The system shown comprises a first optical sensor 10, set up for a directional detection of brightness values in a first field of view 12, wherein a signal output of the first optical sensor 10 is connected to a first transimpedance amplifier 14 with an adjustable gain Viva" set up so as to generate an output signal II proportional to a brightness in the first field of view 12; a second optical sensor 20 set up for a directional detection of brightness values in a second field of view 22, wherein the first field of view 12 and the second field of view 12 are aligned essentially parallel to each other, at a fixed separation distance B from each other, wherein a signal output of the second optical sensor 20 is connected to a second transimpedance amplifier 24 with an adjustable gain V2var, set up so as to generate an output signal 12 proportional to a brightness in the second field of view 22; a means for control 40 of the first and the second optical sensors 10, 20, wherein the means for control 40 continuously alternates between an adjustment phase and a measurement phase, wherein in the adjustment phase an adjustment takes place of the gains Vlvar, V2var, of the first and second transimpedance amplifiers 14, 24 respectively, to a defined preferred operating range corresponding to a current brightness in the first and second fields of view 12, 22, wherein in the subsequent measurement phase, with the previously adjusted gains Viset, V2", of the first and the second transimpedance amplifiers 14, 24, a measurement of the time profiles of the brightness in the first and second fields of view 12, 22 is performed for determining the velocity of passing objects 0.

The first and second optical sensors 10, 20 can be arranged at a fixed separation distance from each other by means of a measuring beam 30 (or a support or holder). However, such a measuring beam 30 is not necessary if the separation distance between the first and second field of view is determined in another manner, and is known to the system for further evaluation. In particular, the first and second optical sensors 10, 20 can thus also be designed to be positioned independently of each other, so that they can also be arranged, for example, in/at different windows, or window openings, within a room or a flat.

The system as shown further comprises a means for evaluation 50, designed so as to identify or classify an object passage P in the time profiles, from the time profiles of the brightness in the first and second fields of view 12, 22, measured in a measurement phase on the basis of analytical and/or statistical methods using defined parameters. The means for evaluation 50 can further be designed so as to determine the velocity of the object 0, for an identified or classified object passage P, on the basis of a polynomial fitting at each of the extreme points belonging to the identified or classified object passage P in the measured time profiles of the brightness in the first and second fields of view 12, 22, so as to determine analytically the exact time location of the extreme points by forming the derivative of the respective polynomial fitting, and so as to determine the velocity of the object 0 with the time difference resulting from the location of the extreme points by difference formation, together with the separation distance B between the first and second fields of view 12, 22.

As shown, the system preferably comprises a means for location determination 60, set up so as to determine a tilt of the connecting line between the first and the second fields of view 12, 22 with respect to a predetermined axis of the object passages P, and so as to transfer it to the means for evaluation 50 as a correction to the separation distance B between the first and the second fields of view 12, 22.

Figure 2 shows a schematic illustration of an optical sensor with a transimpedance amplifier (TIA), a "sample-and-hold" element (S&H) and an instrumentation amplifier (INA). The sensitivity of the measurement, and thus the gain Vva, of the TIA, is adjustable. When light is incident on the optical sensor (e.g. a photodiode), an output signal 11 proportional to the brightness of the incident light is output as a signal strength from the TIA. The "sample-and-hold" element can be used so as to subtract a background brightness from the output signal in the measurements (a background brightness correction). Accordingly, the "sample-and-hold" element has an input for defining a specific "hold" value. In particular, an average brightness determined during an adjustment phase, a maximum brightness, or another brightness value derived from the adjustment phase, can be used as a control signal, as an approximate background brightness value for a subsequent measurement phase. During a measurement phase, the "hold" value preferably remains unchanged. By means of the downstream INA, a difference signal, largely corrected for any background brightness, can thus be derived and stored.

For example, the photocurrent of a photodiode, as the optical sensor 10, 20, can be variably amplified by way of a transimpedance amplifier (TIA). In the adjustment phase, the amplified photocurrent can be used directly as a measurement signal for sensitivity adjustment (that is to say, for adjusting the adjustable gain Vivar, V2v21) of the TIA to the current environmental situation. In parallel, this signal can be applied to the input of a "sample-and-hold" (S&H) element. In the subsequent measurement phase, the connection between the transimpedance amplifier and the "sample-and-hold" element can be interrupted. The measurement signal of the transimpedance amplifier, and the current "hold" value of the "sample-and-hold" element, can then be subtracted from each other, and amplified in the means for evaluation, in the simplest case by means of a difference or instrumentation amplifier (INA), and can then result in the actual sensor signal, which can be recorded over the duration of a measurement phase. At the end of the measurement phase, the system can switch back to the adjustment phase, by way of the means for control, and the process as described can be repeated.

The data flows of two such optical sensors, arranged at a known separation distance, can be collected by the means for evaluation in the course of the measurement phase. Subsequently, various evaluation methods can be applied to each of them for executing an object recognition, that is to say, an identification or classification of object passages. If corresponding events are recognised as valid, the means for evaluation can determine the time difference between the features belonging to an object passage in the time profiles.

Figure 3 shows an illustration of the relative frequency of half-widths of the cross-correlation function, using an exemplary data set. The evaluation shown is based on a data set with approx. 650 individual measurements. Using these statistics, statements can be made about the distribution of the velocities of the objects, and the permissible evaluation range. However, such an evaluation can also contribute to an improved identification or classification of objects, with the application of appropriate statistical methods. The bars shown in white show the distribution of half-widths, which can be assigned to the distance distribution of the extreme values between the individual profiles. The half-widths belonging to this range are suitable for further processing by means of a polynomial fitting, and correspond to real object passages. The half-widths in the region of the columns marked in grey, on the other hand, are regarded as invalid and are discarded, since in these cases a largely time-synchronous brightness modulation occurred in both fields of view. Thus, in this example, the statistical evaluation shown can be automatically used to sort out half-widths from a threshold value of 1.5 seconds onwards.

Figure 4 shows normalised cross-correlation functions of various signal profiles (a-c) and related signal profiles (d-f). In particular, a) shows a function profile with a large half-width, and b) shows a function profile with a smaller half-width. In c) an example of the occurrence of a plurality of maxima and minima is shown. The other figures show examples of real-time signal profiles of the brightness at the first and second optical sensors (sensor 1, sensor 2), wherein in d) is illustrated a uniformly fluctuating ambient light applied to both optical sensors (due, for example, to clouds, the basis of calculation for Figure 4a), in e) is illustrated the drive-by of a car (a target object, the basis of calculation for Figure 4b), and in f) is illustrated an alteration in brightness occurring at only one sensor (caused for example by a reflection, or an asymmetrical headlight; the basis of calculation for Figure 4c).

Figure 5 shows a schematic illustration of a polynomial fitting of extreme values. The adjustment is exemplarily demonstrated using the time profiles of the brightness at the first and second optical sensor (sensor 1, sensor 2) shown in Fig. 4e) for the drive-by, or passing, of a car (a target object). Here a) shows the entire view of the signal profiles, and b) shows an enlarged view of the profiles around the extreme points. The polynomial fitting of the extreme values is performed in a narrow range around the respective extreme values. In particular, in b) a good agreement between the two polynomial fittings and the original measurement profiles is clearly visible. The exact location of the extreme points can then be determined by simple analytical derivation and calculation of the null points of the appropriately fitted polynomial.

List of reference symbols First optical sensor

12 First field of view

14 First transimpedance amplifier Second optical sensor

22 Second field of view

24 Second transimpedance amplifier Measuring beam Means for control Means for evaluation Means for determining location O Object B Separation distance P Object passage

Claims

Claims 1. A system for determining the velocity of passing objects (0), comprising: A first optical sensor (10), set up for a directional detection of brightness values in a first field of view (12), wherein a signal output of the first optical sensor (10) is connected to a first transimpedance amplifier (14) with an adjustable gain Viyar, set up so as to generate an output signal 11, proportional to a brightness in the first field of view (12); A second optical sensor (20), set up for a directional detection of brightness values in a second field of view (22), wherein the first field of view (12) and the second field of view (12) are aligned essentially parallel to each other at a fixed separation distance (B) from each other, wherein a signal output of the second optical sensor (20) is connected to a second transimpedance amplifier (24) with an adjustable gain V2,,a,, set up so as to generate an output signal 12, proportional to a brightness in the second field of view (22); A means for control (40) of the first and the second optical sensors (10, 20), wherein the means for control (40) continuously alternates between an adjustment phase and a measurement phase, wherein in the adjustment phase an adjustment takes place of the gains Vlvar, V2var of the first and second transimpedance amplifiers (14, 24), in each case to a defined preferred operating range corresponding to a current brightness in the first and second fields of view (12, 22), wherein in the subsequent measurement phase, with the previously adjusted gains Vlset, V25e1 of the first and of the second transimpedance amplifiers (14, 24), a measurement is performed of the time characteristics of the brightness in the first and second fields of view (12, 22) for determining the velocity of passing objects (0).
2. The system according to Claim 1, wherein the duration of an adjustment phase is shorter than the duration of the subsequent measurement phase, wherein the duration of the adjustment phases and the duration of the measurement phases are in each case constant, or can be parametrically adjusted.
3. The system according to Claim 1 or 2, further comprising: A means for evaluation (50), designed so as to identify or classify an object passage (P) in the time profiles, from the time profiles of the brightness in the first and second fields of view (12, 22), measured in a measurement phase on the basis of analytical and/or statistical methods using defined parameters.
4. The system according to Claim 3, wherein the analytical and/or statistical methods for identifying or classifying object passages (P) use at least one criterion from the following group: Alteration of the difference signal between the measured time profiles of the brightness in the first and second fields of view (12, 22) by a defined amount; Mean values and/or standard deviations, or quantities calculated therefrom, in the difference signal between the measured time profiles of the current brightness in the first and second fields of view (12, 22); Differences between arithmetic mean values, the confidence intervals, or the difference of the standard deviations of mean values in the measured time profiles of brightness in the first and second fields of view (12, 22); A half-width in a correlation function between the measured time profiles of the brightness in the first and second fields of view (12, 22); A number of maxima and/or minima in a correlation function between the measured profiles of the brightness in the first and second fields of view (12, 22).
5. The system according to Claim 3 or 4, wherein the means for evaluation (50) for an identified or classified object passage (P) determine the velocity of the object (0) on the basis in each case of a polynomial fitting at the extreme points belonging to the identified or classified object passage (P) in the measured time profiles of the brightness in the first and second fields of view (12, 22), analytically determine the exact time location of the extreme points by forming the derivative of the respective polynomial fitting, and determine the velocity of the object (0) with the time difference resulting from the location of the extreme points by difference formation, together with the separation distance (B) between the first and second fields of view (12, 22).
6. The system according to one of the preceding claims, further comprising a means for determining location (60), set up so as to determine a tilt of the connecting line between the first and the second fields of view (12, 22) with respect to a predetermined axis of the object passages (P), and to transfer it to the means for evaluation (50) as a correction to the separation distance (B) between the first and the second field of view (12, 22).
7. A method for determining the velocity of passing objects (0), comprising: Provision of a system for determining the velocity of objects (0) according to one of the preceding claims; Continuous alternation between an adjustment phase and a measurement phase of the system, wherein in the adjustment phase an adjustment is performed of the gains Vlvar, V2var of the first and second transimpedance amplifiers (14, 24) respectively to a defined preferred operating range corresponding to a current brightness in the first and second fields of view (12, 22), wherein, in the subsequent measurement phase, a continuous measurement of the profiles of the current brightness in the first and second fields of view (12, 22) is performed with the previously adjusted gains Vlset, V25et of the first -20 -and second transimpedance amplifiers (14, 24) for determining the velocity on the basis of identified or classified object passages (P).
8. The method according to Claim 7, further comprising: Identification or classification of object passages (P) from the profiles of the current brightness, measured in a measurement phase, in the first and second fields of view (12, 22) on the basis of analytical and/or statistical methods using defined parameters.
9. The method according to Claim 8, wherein the analytical and/or statistical methods for the identification or classification of object passages (P) use at least one criterion from the following group: Alteration of the difference signal between the measured profiles of the brightness in the first and second fields of view (12, 22) by a defined amount; Mean values and/or standard deviations, or quantities calculated therefrom, in the difference signal between the measured profiles of brightness in the first and second fields of view(12, 22); Differences between arithmetic mean values, the confidence intervals, or the difference of standard deviations of mean values in the measured profiles of brightness in the first and second fields of view (12, 22); A half-width in a correlation function between the measured profiles of brightness in the first and second fields of view (12, 22); A number of maxima and/or minima in a correlation function between the measured profiles of the brightness in the first and second fields of view (12, 22).
10. The method according to Claim 8 or 9, further comprising: Polynomial fittings respectively at each of the extreme points belonging to the determined or classified object passage (P) in the measured profiles of the current brightness in the first and second fields of view (12, 22); Analytical determination of the exact time location of the extreme points by formation of the derivative of the respective polynomial fitting; Determination of the velocity of the object (0) with the time difference resulting from the time location of the extreme points by difference formation, together with the separation distance (B), wherein a tilting of the connecting line between the first and the second fields of view (12, 22) with respect to a predetermined axis of the object passages (P) can be taken into account as a correction to the separation distance (B) between the first and the second fields of view (12, 22).